No — not in the way that matters for a business. AI can weigh more variables, apply rules more consistently, and surface options no human would reach, but it cannot hold the value trade-off that makes a choice ethical, and it cannot answer for the outcome afterward. A system produces more or less ethical results depending on how it is built and governed; the moral decision, and the accountability for it, stay with the people who deploy it.
In a recent DBA seminar on business ethics and governance, the most useful question in the room was not whether AI is biased — that is now assumed — but whether a leader can legitimately delegate a moral call to a system that has no stake in the result. Strip the hype away and the answer reorganizes how executives should position AI in the decision chain.
Can an AI actually make an ethical decision?
No. It can compute one, which is not the same thing. An ethical decision requires weighing competing values — fairness against speed, loyalty against profit, near-term harm against long-term good — and being willing to own the trade-off. As organizational scholars have argued, a system would need genuine moral agency, durable intentions to act well under uncertainty, and the capacity to take responsibility before it could be considered an ethical decision-maker in its own right. AI has none of these. What it has is pattern recognition over past human judgments. That makes it a powerful mirror: it reflects our data, our priorities, and our blind spots back at us at scale. A mirror can sharpen a decision. It cannot make one.
What happens to a moral choice when you route it through AI?
The value judgment does not disappear — it moves upstream and goes quiet. Every model encodes choices: what data it learned from, what objective it optimizes, where the approval threshold sits, how the prompt is framed. Those are moral decisions made by people, often months earlier and several teams away. By the time a clean, confident recommendation reaches the executive, the trade-off is invisible. This is judgment laundering: a contestable human value choice enters the system, gets processed, and exits looking like a neutral, technical output. “The model recommended it” quietly replaces “I decided, and here is why.” The danger is not that AI is unethical — it is that AI makes the ethics unexaminableunless you deliberately surface them.
Why isn’t “the AI recommended it” an ethical or legal defense?
Because the ought cannot be outsourced. Vendors disclaim liability for how their models are used; responsibility lands on the organization that deployed the system. Regulators have made the same move. The EU AI Act requires meaningful human oversight of high-risk systems, and in the United States, courts and commentators increasingly read board-level AI oversight into existing fiduciary duties — Delaware’s standards now expect directors to govern AI risk rather than defer to it. Recent 2026 analysis of AI in business decision-making warns that algorithmic delegation erodes corporate accountability precisely because it lets consequential decisions be made with no identifiable human author. The recommendation may come from a machine; the answer for it does not.
Does relying on AI erode ethical reasoning over time?
It can, and this is the quieter risk. Moral judgment is a practiced capability — leaders get better at hard calls by making them. When the system supplies the answer and the human supplies only a signature, the reasoning muscle atrophies. Over time an organization can lose the ability to even recognize a moral trade-off, not just resolve one. This is also where “ethics washing” takes hold: AI ethics policies, principles, and committees that perform ethical seriousness without ever exercising it. Securities regulators now treat AI-washing — overstated or hollow claims about responsible AI — as a real disclosure and compliance risk, not a public-relations footnote.
How should boards and executives govern AI so ethics stay human?
Keep humans in the value-setting loop, not only the output-review loop. Reviewing a recommendation after the fact is too late; the ethics were decided upstream. Concretely: require that consequential AI decisions name the trade-offs they encode and the human who owns them; build an AI ethics or governance committee with real audit authority rather than advisory cover; and raise AI literacy at board level, because directors cannot oversee what they cannot interrogate. The gap is documented — in 2026 roughly two-thirds of directors report using AI for board work, but only about a fifth have governance processes for it, and most boards are judged to have only limited AI expertise. The regional pattern sharpens the point. In the Gulf, where status-driven adoption is fast and growth-oriented, ethical discipline is the step rapid deployment is most tempted to skip. In the US, scale and efficiency push toward deployment-first. In Central Europe, cost discipline can crowd out oversight investment. In each case, governance is not a brake on AI — it is the condition for using it without laundering away your own judgment.
Frequently asked questions
Can AI be programmed to be ethical?
It can be designed to produce more ethical outcomes — bias audits, fairness constraints, transparency requirements — but that is shaping behavior, not granting moral agency. The system still holds no values of its own and cannot take responsibility, so the ethical ownership remains human.
What is “ethics washing” in AI?
Ethics washing is superficial ethical signaling — principles, pledges, or committees — without substantive commitment or action, often used to deflect scrutiny or pre-empt regulation. In 2026 it carries real exposure: regulators increasingly treat overstated responsible-AI claims as a compliance and disclosure issue.
Who is responsible if an AI makes an unethical decision?
The organization that deploys it. Model providers generally disclaim liability for downstream use, and oversight obligations attach to the deployer. Accountability sits with the leaders and the board, not the tool.
Should a company have an AI ethics committee?
Increasingly, yes — and increasingly it is read as part of fiduciary duty. But the committee earns its place only if it has audit authority and shapes the value choices embedded in systems, rather than ratifying decisions already made. A committee that only signs off is governance theater.
